adw-bootstrap

# AI Developer Workflows Bootstrap Skill

## Mission

Bootstrap **AI Developer Workflows (ADWs)** infrastructure that enables programmatic agent orchestration in any codebase. Transform a regular project into one where AI agents can be invoked programmatically to plan, implement, test, and deploy features.

## What ADWs Enable

After setup, developers can:
- **Execute prompts programmatically**: `./adws/adw_prompt.py "implement feature X"`
- **Use reusable templates**: `./adws/adw_slash_command.py /chore "task"`
- **Orchestrate multi-phase workflows**: Plan → Implement → Test → Deploy
- **Track agent behavior**: Structured outputs in `agents/{id}/` for debugging
- **Scale compute**: Run multiple agents in parallel for complex tasks

## Core Philosophy: Intelligence Over Templating

**You are NOT executing a rigid template substitution.**

You will:
1. Read working reference implementations
2. Understand the patterns they demonstrate
3. Analyze the target project's structure and conventions
4. Intelligently adapt the references to fit the target
5. Make contextual decisions based on project needs

**Use your reasoning.** Handle novel structures, mixed languages, and edge cases that no template could anticipate.

## Two-Layer Architecture

ADWs create a **two-layer architecture**:

1. **Agentic Layer** (`adws/`, `.claude/`, `specs/`) - Templates engineering patterns, teaches agents how to operate
2. **Application Layer** (`apps/`, `src/`, etc.) - The actual application code that agents operate on

The agentic layer wraps the application layer, providing a programmatic interface for AI-driven development.

## Progressive Enhancement Model

Setup happens in phases based on project needs:

- **Minimal** (Always): Core subprocess execution, basic prompts, essential commands
- **Enhanced** (Recommended for dev projects): SDK support, compound workflows, richer commands
- **Scaled** (Production/teams): State management, triggers, testing, worktree isolation

## IMPORTANT: U